A systematic review and meta-analysis of randomized controlled trials found that substituting soymilk for cow's milk was associated with improvements in several cardiometabolic risk factors.

In the study, published in BMC Medicine, investigators analyzed data from 17 trials involving 504 adult participants. The investigators searched MEDLINE, Embase, and the Cochrane Central Register of Controlled Trials databases through June 2024. Inclusion criteria were randomized controlled trials in adults lasting at least 3 weeks that compared soymilk with cow's milk and reported on cardiometabolic outcomes.

The meta-analysis included trials with a median duration of 4 weeks (range = 4–16 weeks) and a median soymilk dose of 500 mL/day (range = 240–1,000 mL/day), containing 22 g/day of soy protein (range = 2.5–70 g/day). For sweetened soymilk, the median added sugar content was 17.2 g/day (range = 4.0–32 g/day).

The analysis used random-effects models to pool data across studies. Heterogeneity was assessed using I2 statistics and explored through sensitivity analyses. The certainty of evidence was evaluated using the GRADE approach.

Study characteristics included:

• 17 trials (10 crossover and seven parallel design)
• Median 25 participants per trial (range = 7–60)
• Median age: 48.5 years (range = 20–70)
• Median body mass index (BMI): 27.9 kg/m2 (range = 20–31.1)
• Participants included those with hypercholesterolemia (four trials), overweight/obesity (four trials), type 2 diabetes (two trials), hypertension (one trial), rheumatoid arthritis (one trial), and healthy individuals (three trials)
• 392 female and 112 male participants.

The median cow's milk dose was also 500 mL/day (range = 236–1,000 mL/day), containing 24 g/day of milk protein (range = 3.3–70 g/day) and 24 g/day of total sugars as lactose (range = 11.5–49.2 g/day).

Among the key findings were:

• Reductions in non–high-density lipoprotein (HDL) cholesterol (–0.26 mmol/L, 95% confidence interval [CI] = –0.43 to –0.10 mmol/L, P < .001)
• Decreased systolic blood pressure (–8.00 mmHg, 95% CI = –14.89 to –1.11 mmHg, P = .023)
• Lowered diastolic blood pressure (–4.74 mmHg, 95% CI = –9.17 to –0.31 mmHg, P = .036)
• Reductions in low-density lipoprotein (LDL) cholesterol (–0.19 mmol/L, 95% CI = –0.29 to –0.09 mmol/L, P < .001)
• Decreased C-reactive protein (–0.81 mg/L, 95% CI = –1.26 to –0.37 mg/L, P < .001)
• An increase in HDL cholesterol (0.05 mmol/L, 95% CI = 0.00 to 0.09 mmol/L, P = .036)
• No significant effects on triglycerides, fasting glucose, or body weight.

The analysis found no statistically significant differences between sweetened and unsweetened soymilk.

Subgroup and sensitivity analyses found no significant effect modification by factors such as health status, BMI, age, or soy protein dose. The analysis noted trends toward greater LDL cholesterol reduction in participants with hypercholesterolemia, higher baseline LDL cholesterol, and higher soy protein doses (> 25 g/day).

The certainty of evidence was rated as high for LDL cholesterol and non-HDL cholesterol outcomes. It was moderate for blood pressure, C-reactive protein, and HDL cholesterol because of imprecision in the pooled estimates. The evidence was low for HbA1c, 2-h plasma glucose, creatinine, eGFR, ALT, and AST because of indirectness and imprecision.

Dose-response analyses were performed for LDL cholesterol, HDL cholesterol, non-HDL cholesterol, triglycerides, body weight, and BMI. A positive linear dose-response was observed for triglycerides (P = .038), but no significant dose-response relationships were found for other outcomes.

Publication bias was assessed for LDL cholesterol using contour-enhanced funnel plots and formal testing (Begg's test P = .721, Egger's test P = .856), with no evidence of asymmetry detected.

Adverse events were reported in nine trials. Digestive difficulties related to soymilk consumption led to 1 to 2 withdrawals in three of the trials, while two trials reported similar issues with cow's milk consumption. Four trials indicated that most participants found both soymilk and cow's milk acceptable and tolerable.

Limitations included imprecision in estimates for several outcomes and indirectness for some markers with limited available data. The investigators noted a need for more research on glycemic control, kidney function, and nonalcoholic fatty liver disease outcomes.

This study provided data on the effects of substituting soymilk for cow's milk on cardiometabolic risk factors. The findings may have implications for dietary guidelines and food labeling policies related to plant-based alternatives.

The authors declared having no competing interests.